On Sep 1, 2016, at 08:42, Richard Hamilton <user-af55987f6d56@xymon.invalid> wrote:
It's a useful enough possibility...and we all use workarounds from time to time. But as time permits, the real cause of the problem should be found and fixed; and workarounds should not become permanent. I'd worry that a monitoring tool being used to implement workarounds could make that too tempting.
On Thu, Sep 1, 2016 at 8:01 AM, Henrik Størner <user-ce4a2c883f75@xymon.invalid> wrote:
Den 01-09-2016 07:33, Bruce Ferrell skrev:
I've never ever seen it written up, and I've been a user since the
days when xymon was BigBrother (yes, I know they aren't common code...
xymon IS the spiritual descendant however)
Recently, I started monitoring a remote service that was failing
regularly. It would send me an alert and I'd go fix the service. I
got tired of having to do the restarts so I
looked a bit more into alerts.cfg. Yes, I can send an alert via a
script (do that all the time for sms)... Wait... Can that script do
anything else? Well, I'll be! I wrote one
to ssh into the offending system (key based authentication) and
perform simple diags, collects relevant logs then restart the downed
service.
I have seen a number of times, "xymon/bigbrother doesn't restart things".
Thoughts? Concerns?
Sure, you can do that if it suits your way of working. And I can certainly see why it would be nice to avoid restarting the same service again and again.
The reason that Xymon doesn't do that "out of the box" is this: Xymon has always been a "watch, but don't act" tool. And that is an inheritance from the Big Brother days.
Regards,
Henrik